N3-methyladenine (3mA) and 1,N6-ethenoadenine (eA) are two DNA base modifications produced from exposure to environmental genotoxic agents, cellular metabolites, and anti-cancer drugs. 3mA lesions are highly cytotoxic owing to their inhibition of DNA synthesis by polymerases, and this cytotoxicity is a rationale for the use of alkylating agents in cancer chemotherapy. eA, which is associated with chronic inflammatory conditions, is highly mutagenic and can lead to genomic instability and cancer. Two different, partially redundant enzymatic activities have evolved for these two specific lesions: i) oxidative demethylation by DNA dioxygenases and ii) base excision repair by DNA glycosylases. The precise determinants for the specificity and catalysis of these enzymes toward 3mA and eA remain unclear. We seek to fill this critical gap in knowledge by a unique integration of directed evolution and structural biology methods in order to obtain a comprehensive mechanistic understanding of 3mA and eA selection and catalysis by human ABH2 dioxygenase (Aim 1) and the yeast family of MAG 3mA glycosylases (Aim 2). This work capitalizes on the convergent evolution observed between the two repair systems, and is based on our preliminary results that have identified ABH2 mutants with the capacity to protect cells from 3mA toxicity. We will test the hypothesis that ABH2 repair of 3mA, unlike that of other known substrates, involves excision and further processing by base excision repair. Our general approach for each aim is to i) identify residues important for substrate discrimination using directed evolution under selective alkylation pressure, ii) determine crystal structures of ABH2 and MAG proteins in complex with 3mA- and eA-DNA, and iii) test the contribution of individual residues toward 3mA and eA specificity and repair. These studies will provide novel insight into how these enzymes determine the fate of cytotoxic and mutagenic lesions toward a particular repair pathway. In addition, in Aim 1 we probe the translational implications of our ABH2 mutants for cancer treatment with methylating agents using a mouse erythroleukemia (MEL) cell tissue culture model. Our studies have at least three direct clinical implications. First, etheno-DNA adducts likely play a role in the etiology of cancer associated with chronic inflammation, and thus results on eA repair may provide new ways to determine the risk of cancer in patients suffering from chronic inflammatory conditions. Second, our 3mA-protecting ABH2 mutants have direct implications for understanding the origins of resistance to therapy with methylating agents in tumors and for the design of new chemotherapeutic approaches involving bone marrow protection. Third, our structure-function studies on 3mA glycosylase repair are a necessary first step for the design of small molecule inhibitors as a way to enhance the cytotoxicity of methylating agents.